1. Introduction
In recent years, there has been an increased interest in body-centric sensor and communication systems due to their wide range of applications, which aim to enhance quality of life by improving sensing, detection, and monitoring capabilities [
1,
2,
3]. To meet the requirements of these applications, wearable antennas must be portable, flexible, and capable of integrating into clothing or uniforms. As a result, these antennas must simultaneously meet technical requirements, safety regulations, aesthetic standards, and the demands of wearable applications. In general, body-centric wireless systems are divided into three groups: off-body, on-body, and in-body systems, each with its own design challenges [
4]. The on-body and off-body modes of operation are mainly related to the communication of different wearable sensors with a central node to aggregate measured data and transmit them to the access point (see
Figure 1). In the in-body mode of operation, in addition to a reliable communication link between implanted and external monitoring devices, wireless power transfer (WPT) systems are often required to solve the problem of powering the implanted devices. A wide range of applications has been proposed, ranging from the ability to monitor bio signals and bio parameters to the capacity to interact with the environment. This continuous stream of information about a person’s health status can provide valuable medical insights, in addition to enabling the tracking of fitness levels in healthy individuals. Such systems can also be applied in emergency operations, such as firefighting operations. Communication capabilities can find use in contexts such as the implementation of safety vest systems necessary in environments where human–robot collaboration is essential, such as in automated warehouses [
5].
From the user’s perspective, factors like flexibility and comfort are of concern. Therefore, antennas should be seamlessly integrated into garments. Two approaches exist for creating such antennas: one involves using non-textile materials, like flexible copper antennas, customized ornamental antennas, or plaster-like antennas designed for direct application on the skin, and the alternative approach is to employ textile antennas [
6]. Since the antennas under consideration are mostly integrated into clothing, designs made of conductive textiles are preferred, as they allow the wearer freedom in daily activities while enabling connectivity between devices.
The vast majority of existing wearable antennas are dipoles, monopoles, planar inverted Fs (PIFAs), microstrip patches, and planar low-profile antennas in general [
2,
7,
8,
9,
10]. However, other types of antennas can also be adapted to be suitable for wearable applications. The aim of this work is to further investigate the properties of wearable realization of a slotted waveguide antenna [
11] made using conductive fabric. A waveguide slot antenna made of a textile-filled waveguide is presented in [
12], but the waveguide walls are made of copper foil, which is impractical for wearable applications. The application of surface-integrated waveguides (SIWs), adapted for textiles as textile-integrated waveguides (TIWs) [
13,
14,
15,
16,
17], has also been proposed. A TIW comprises three layers: the top and bottom conductive layers are typically constructed from conductive fabric, while the central dielectric layer typically consists of traditional materials such as wool felt or polyester-based textile. The creation of the narrow waveguide walls is adapted to textile technology, often achieved using conductive thread or eyelets. TIWs are well-suited for higher frequencies, e.g., in the millimeter-wave frequency range. However, for lower frequencies, it is necessary to employ multiple layers of textile, potentially resulting in a bulkier design. Moreover, due to the utilization of conductive thread or eyelets in achieving narrow waveguide walls, incorporating slots onto them becomes unfeasible. This limitation becomes especially pronounced when developing waveguide antennas for on-body modes of operation (as discussed in detail in
Section 4).
In previous work [
18,
19], we proposed a waveguide antenna that was fully realized using conductive fabric and a sewing manufacturing procedure, which is a natural way of making textile objects. This allows for easy installation of the proposed antenna into a jacket, smart vest, belt, or other clothing item. We theoretically demonstrated the robustness of the proposed solution in all three (off-body, on-body, and in-body) communication scenarios. The experimental realization was focused on the radiation properties and inherent losses of the textile slotted waveguide array; therefore, we fed the developed antennas using a commercial waveguide-to-coax adapter, which should be replaced in the final design. Furthermore, the material used for mold realization was ridged and quite thick; thus, improvements are possible in order to achieve a fully wearable and bendable solution. Therefore, the goal of this paper is to discuss different aspects of the realization of textile slotted waveguide antennas, including the design procedure, suitable feeding structure (coax-to-waveguide transitions with emphasis on simplicity of design), mold selection to enable the creation of a thin and flexible antenna structure, etc. Various feeding structures were initially considered in [
20]. However, this paper is focused on presenting a systematic design approach and discussing the antennas that were subsequently realized based on this approach in detail.
This paper is organized as follows.
Section 2 presents the realization and basic properties of textile slotted waveguide antennas with an emphasis on different modes of operation: off-body, on-body, and in-body scenarios. The problem of the feeding structure is discussed in
Section 3, and three types of coax-to-waveguide transitions are proposed and characterized. Six textile antenna prototypes are manufactured and experimentally characterized, and a comparison of measured and calculated results is presented in
Section 4. An approach to estimate the coupling level between wearable antennas in the on-body mode of operation is also discussed. Finally, conclusions are presented in
Section 5.
2. Realization of Textile Slotted Waveguide Antenna
Body-centric wireless communication systems have many potential applications. Antennas used in such systems are typically classified into three scenarios based on the communication channel: on-body, off-body, and in-body mode of operation.
Figure 1 illustrates the typical applications for each scenario. In on-body communication systems, antennas are designed to enable the propagation of electromagnetic waves along the surface of the human body, allowing for seamless communication between sensors and central nodes (gateways). Conversely, off-body antennas require a wide beam around the human body to establish communication between a central node on the human body and access points positioned at arbitrary locations. Lastly, in-body antennas are optimized to efficiently radiate electromagnetic waves into the human body, enabling reliable communication between central nodes and implantable sensors, as well as facilitating power transfer to these devices. Therefore, the actual realization of a slotted waveguide antenna depends on the specific application.
As examples, two different antenna designs are presented in
Figure 2 for off-body and on-body types of communication. In the first case, the array contains three slots that are half the guided wavelength apart and approximately half a wavelength long. The waveguide is short-circuited at a distance of three-quarters of the guided wavelength from the center of the third slot. By doing so, each slot radiates a portion of both a forward and backward propagating wave, resulting in increased radiated EM power per slot. Note that most of the energy is radiated out from the body; thus, the regulations concerning the body specific absorption rate (SAR) values are easily satisfied.
For the on-body mode of operation, the slots are cut into the narrow waveguide walls, resulting in the launch of the normally polarized EM wave along the body. Additionally, the penetration depth of EM fields into the human body is very small, fulfilling the safety requirements. For both cases, the calculated SAR values and the corresponding maximum input antenna powers are given in [
18].
In order to investigate the properties of the proposed textile slotted waveguide antenna, we designed and experimentally verified several textile waveguide antennas operating in the 5.8 GHz ISM band. A schematic representation of the considered textile antennas is shown in
Figure 2. In principle, the antenna is made of a piece of waveguide on which slots are cut out and which is short-circuited at the end. The waveguide walls were made of conductive textile, and all connections of the walls were realized by a classical sewing procedure. The radiating slots were cut out, and the borders were sewn to fix the dimensions and prevent tearing. The sewn slotted waveguide, with a shape resembling that of a sock, was pulled over a mold (i.e., over an appropriate supporting structure) in order to keep the desired cross section of the antenna.
For the conductive fabric, we selected Shieldex®Nora Dell No.: 1401101S80 [
21]. This conductive textile is made of Ni/Cu/Ag-plated polyamide fabric with an average surface resistivity of 0.009
and a thickness of
mm ±
. To create a mold, two materials were considered: rigid molds made from styrodur planar sheets with a permittivity of
[
18] and bendable molds made using Cuming Microwave C-FOAM PF-4 sheets with a permittivity of
[
22]. The use of a bendable mold offers several advantages compared to the use of a stiff Styrofoam mold. The bendable mold is thinner, making it more suitable for integration into clothing and enabling the realization of flexible textile waveguide antennas. Therefore, it is important to further investigate and experimentally characterize the performance of textile waveguide antennas with a thin foam-based mold. This will help to determine the feasibility and potential benefits of using a thin bendable mold for antenna integration into clothing and wearable applications.
3. Feeding Possibilities
The design of the considered waveguide antennas is divided into two parts, where the dimensions are separately optimized. First, the transition from an SMA coaxial connector to a rectangular waveguide is designed. Then, in the second step, a slotted waveguide array is constructed. In the first step, the waveguide is loaded with the matched waveguide port, ensuring that all the power entering the waveguide is transmitted to the waveguide port. In the second step, the antenna is excited with a waveguide-dominant mode using a waveguide port in a general electromagnetic solver (CST in our case [
23]). In the final step, the two designed structures are joined together, and the antenna dimensions are fine-tuned. One could also design the feeding structure and the antenna part together. However, in such a case, there is a non-negligible possibility of designing the antenna as a resonator loaded with radiating slots (note that the considered waveguide section is short-circuited at both sides; see
Figure 2), leading to narrow operating bandwidth design.
In this section, we discuss three different types of feeding transitions, as shown in
Figure 3:
Transition A: Top- or bottom-mounted coax-to-waveguide transition;
Transition B: Edge-mounted coax-to-waveguide transition;
Transition C: Microstrip line-to-waveguide transition.
The width and height of the waveguide are denoted as
a and
b, respectively, while the permittivity of the waveguide filling is denoted as
. Two types of molds were considered: a rigid mold with a height of
mm (made from Styrodur) and a bendable mold with a height of
mm (made from foam PF-4). Final waveguide dimensions are provided in
Table 1. Transitions A and C were later experimentally realized and integrated with the antenna to characterize their properties and their influence on the radiation properties of the antenna.
Transitions A and B were designed using a bottom- and edge-mounted SMA connector, respectively (Radiall R125.414.000 was used in all prototypes), creating a structure similar to that of a waveguide-to-coax adapter. However, unlike high-quality adapters, our focus was on simplicity of design, i.e., the goal was to design a “good enough” transition covering at least the 5.8 GHz ISM band (5.725–5.875 GHz). Therefore, related to transition A, only two parameters were varied to obtain an optimal design: the length of the pin (the diameter was equal to the diameter of the inner conductor of the coaxial connector, i.e., 1.28 mm) and the distance from the short circuit. In the case of a thinner waveguide (
b = 6.35 mm), we decided to extend the pin to the opposite side of the waveguide wall and solder it to that textile wall, obtaining good contact. Thus, in the design of a thin waveguide, only one dimension was available for optimization: the distance from the short circuit. Since the capacitive part of the matching circuit cannot be obtained by changing the length of the pin, the only way to add the capacitive part is to place the feeding point quite far away (slightly less than
) from the short-circuit wall where the shorted waveguide section shows a capacitive reactance. Therefore, this distance is now much larger—30.5 mm versus 9.0 mm as in the previous case (all the dimensions are given in
Table 2). One should note that a larger bandwidth can be obtained by employing a shorter pin (like in the case of a thicker waveguide). However, such a design is more sensitive to parameter variation. In
Figure 4, the magnitude of the
parameter is given for both designs. Although the considered structure is very simple, with only one or two degrees of freedom, good impedance matching in the entire 5.8 GHz ISM band is achieved. Of course, allowing additional degrees of freedom for optimization would result in better impedance matching in a larger bandwidth.
Transition B consists of an edge-mounted SMA connector feeding a shorting elbow. There are two possibilities with respect to how to feed the shorting elbow: with a wire or with a strip. In contrast to transition A, which essentially excites the waveguide with an
E field from a monopole, transition B produces excitation through an
H field with a current loop. However, transition B is more complicated than transition A, as it has four degrees of freedom. Due to this complexity, transition B is not experimentally realized or integrated with textile waveguide antennas. Nonetheless, the optimized dimensions for transition B, including both wire and strip-fed shorting elbows, as well as for rigid and bendable molds, are provided in
Table 3.
Figure 5 shows the calculated
parameters of transitions B for both types of molds. As shown in
Figure 5, both wire-fed and strip-fed transitions B provide matching in the
GHz ISM band. In general, a strip-fed shorting elbow tends to provide better impedance matching compared to a wire-fed shorting elbow. This is because the strip provides a larger surface area for the EM wave to couple into the waveguide, resulting in improved power transfer and reduced reflection.
The final considered transition is a microstrip line-to-rectangular waveguide transition, as shown in
Figure 3c. The transition includes an SMA connector, a 50
microstrip transmission line, and a linear microstrip taper. Since in this case, radiation losses are present, a symmetrical two-port cascade network of the waveguide and transitions (as shown in
Figure 3c) was analyzed in order to design the transition. The taper width (
w) was obtained using the CST optimizer tool with the goal of maximizing the
parameter at the central frequency of 5.8 GHz. The taper length should be a multiple of a quarter of a guided wavelength [
24,
25], and the best performance is obtained with tapers that are 3
long. The taper dimensions of transition C obtained by optimization are listed in
Table 4.
Figure 6 shows the simulated
and
parameters of a double-cascaded transition network. Based on the
S parameters shown in
Figure 6, it can be concluded that such a transition radiates a significant amount of power. The two-port cascaded network of the two transitions C exhibits a drop in the
parameter (power transfer) of approximately 3 dB because the dielectric used inside the microstrip and waveguide is foam (with permittivity of
), and the E field does not become trapped below the microstrip and taper lines due to the absence of contrast in permittivity. Using a dielectric with a higher permittivity for transition realization would also imply using the same dielectric as a mold for the textile waveguide antenna. However, a slotted waveguide antenna is, in principle, a leaky wave antenna, and efficient slot radiation requires no permittivity contrast between the waveguide filling and air.